The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Conversion - why the slow development? Audio Interfaces
Old 25th January 2015
  #61
Lives for gear
 
Immersion's Avatar
Since 2008, Antelope has released a few intersting converters. I got the Eclipse converter I think it is probably on of the best out there.
Old 25th January 2015
  #62
Gear Maniac
 

Thanks for addressing that stuff psycho monkey and others. Its now my impression that these new 32 bit dac offerings aren't really much different sounding than other converters.

What exactly are they doing though? Does this mean that instead of the daw changing from 32 bit floating point to 24 bit before it sends off to the converter, the converter itself is changing it directly to analog information? While the sound quality is still mainly dependent on the clocking, power supply, and analog components, do you think that those converter chips (ones claiming 32 bit) will represent better sound in interfaces with lesser builds? Would those converter chips offer any improvement to a well built interface?
Old 26th January 2015
  #63
Lives for gear
 
IanBSC's Avatar
The only '32 bit' chip I know much about is the Sabre 9018 and it is high performance. Merging Horus uses this in their basic interface. As I understand it, only DSP such as filtering is 32 bit. According to Michal Jurewicz of Mytek, the antialias filters sound better at 32 bit. This makes sense to me, if was running a DSP filter in a DAW I would rather it be 32 than 24 bit.

But then this is just the DSP precision in the ADC chip, there are a lot more important factors in ADC design. I've heard really detailed implementations of this chip, some of the most detailed converters I have heard (Mytek Manhattan). But I couldn't tell you if the sound quality in these converters had that much to do with the chip.

32 bit DSP may be a slight improvement, but not a game changer.
Old 26th January 2015
  #64
Quote:
Originally Posted by IanBSC View Post
The only '32 bit' chip I know much about is the Sabre 9018 and it is high performance. Merging Horus uses this in their basic interface. As I understand it, only DSP such as filtering is 32 bit. According to Michal Jurewicz of Mytek, the antialias filters sound better at 32 bit. This makes sense to me, if was running a DSP filter in a DAW I would rather it be 32 than 24 bit.

But then this is just the DSP precision in the ADC chip, there are a lot more important factors in ADC design. I've heard really detailed implementations of this chip, some of the most detailed converters I have heard (Mytek Manhattan). But I couldn't tell you if the sound quality in these converters had that much to do with the chip.

32 bit DSP may be a slight improvement, but not a game changer.
I wouldn't claim to know a lot about stuff this technical, but I'm pretty sure we're now not talking about dynamic range in the captured audio.
Old 26th January 2015
  #65
Deleted 6ccb844
Guest
Quote:
Originally Posted by IanBSC View Post
The only '32 bit' chip I know much about is the Sabre 9018 and it is high performance. Merging Horus uses this in their basic interface. As I understand it, only DSP such as filtering is 32 bit. According to Michal Jurewicz of Mytek, the antialias filters sound better at 32 bit. This makes sense to me, if was running a DSP filter in a DAW I would rather it be 32 than 24 bit.

But then this is just the DSP precision in the ADC chip, there are a lot more important factors in ADC design. I've heard really detailed implementations of this chip, some of the most detailed converters I have heard (Mytek Manhattan). But I couldn't tell you if the sound quality in these converters had that much to do with the chip.

32 bit DSP may be a slight improvement, but not a game changer.
At around 120dB dynamic range (roughly 21-Bit) were approaching the edge of limitations, including our hearing and circuitry.

Quantization errors aren't really a huge problem any more, were on the precipice of it's as good as it gets really and if you look at the likes of the Tascam UH-7000 with a low SNR and 123dB dynamic range ADDA (in principle) for £400.00. High performance convertors are seriously cheap.!

In todays market I believe all mid range and high end converters should be able to compete with Tascam's UH-7000 specs.

What I do find ironic from several shootouts, people seem to prefer the least accurate or "noisy" like the MR816. Many prefer the Burl B2 to the Apogee Symphony, but on paper the Symphony wins out by a fair margin. On the DAC side so does an Apogee Duet, or Tascam UH or Focusrite Forte.

It does beg the question and always has, how important are these specs?

Also I'm really not sure what people are actually looking for, some say transparency but it appears that's not always the case.

So I remain by my original statement, what makes a converter good today is price vs. performance, customer support, features, latency and facilities.
Old 27th January 2015
  #66
Lives for gear
 

there is much more to precision and sound quality than dynamic range, guys. just play a 16bit file v a 24bit file through a -65db noisefloor badass chain and you know what i mean.
Old 28th January 2015
  #67
Gear Maniac
 

i guess it would depend on what you want in your convertor. I think the most detailed and transparent converters make more sense, since there are so many ways you can color and smooth sounds before or after conversion.
Old 28th January 2015
  #68
Lives for gear
 
IanBSC's Avatar
Quote:
Originally Posted by psycho_monkey View Post
I wouldn't claim to know a lot about stuff this technical, but I'm pretty sure we're now not talking about dynamic range in the captured audio.
This is just float-point processing.
Old 28th January 2015
  #69
Lives for gear
 
IanBSC's Avatar
Quote:
Originally Posted by ShadowAMD View Post
At around 120dB dynamic range (roughly 21-Bit) were approaching the edge of limitations, including our hearing and circuitry.

Quantization errors aren't really a huge problem any more, were on the precipice of it's as good as it gets really and if you look at the likes of the Tascam UH-7000 with a low SNR and 123dB dynamic range ADDA (in principle) for £400.00. High performance convertors are seriously cheap.!

In todays market I believe all mid range and high end converters should be able to compete with Tascam's UH-7000 specs.

What I do find ironic from several shootouts, people seem to prefer the least accurate or "noisy" like the MR816. Many prefer the Burl B2 to the Apogee Symphony, but on paper the Symphony wins out by a fair margin. On the DAC side so does an Apogee Duet, or Tascam UH or Focusrite Forte.

It does beg the question and always has, how important are these specs?

Also I'm really not sure what people are actually looking for, some say transparency but it appears that's not always the case.

So I remain by my original statement, what makes a converter good today is price vs. performance, customer support, features, latency and facilities.
I'm convinced that there are really important elements of sound quality that are not measured in basic tech specs. Maybe these are slew rate/impulse response, or as of yet unknown properties of sound. I've definitely heard converters that in terms of dynamic range, distortion and jitter were basically perfection, and yet they seemed to lack something and didn't sound right.

I own a 16x16 Burl Mothership and I can say although it has a little bit of color, it's main benefits seem to be a less veiled sound and greater extension in highs and lows. Somehow it is actually mucking up the sound less than other converters in many respects. It does have a bit less detail and range than other converters I've used, and that is reflected in the specs, but sounds more present. So, whatever it is doing better than other converters, seems to be something we don't know how to measure yet.

I'm a bit of a DSD nut, and by standard measurements it's an inferior format up to 256x, and yet it sounds way more realistic (in most respects) to my ears.
Old 28th January 2015
  #70
Gear Guru
 
kafka's Avatar
So what does the world need out of a converter that we don't already have? Your bandwith and S/N ratios are better than any other piece of gear in a studio. You have sample rates that exceed what anybody uses. You can build a better mousetrap, but if your problem is cockroaches, it won't help.
Old 28th January 2015
  #71
Deleted 6ccb844
Guest
@ IanBSC @ kafka

I believe Burl maybe onto something, most still prefer tube amps even though THD @ 5% is pretty high. A lot of high profile audio engineers I've worked with believe it makes it sound better.

Maybe clean and transparent isn't what were looking for? Sure it's great for resolution, but maybe it's at the point where it should be a balancing act as opposed to what specs out the best on paper? I've done A/B tests for new products over the years and the results are often NOT what you'd ever expect.
Old 28th January 2015
  #72
Lives for gear
 
IanBSC's Avatar
Quote:
Originally Posted by ShadowAMD View Post
@ IanBSC @ kafka

I believe Burl maybe onto something, most still prefer tube amps even though THD @ 5% is pretty high. A lot of high profile audio engineers I've worked with believe it makes it sound better.

Maybe clean and transparent isn't what were looking for? Sure it's great for resolution, but maybe it's at the point where it should be a balancing act as opposed to what specs out the best on paper? I've done A/B tests for new products over the years and the results are often NOT what you'd ever expect.
I'm actually surprised the distortion is that high. I am still not sure if the question of transparent vs colored isn't a false one. Some of the cleanest and most transparent converters I have used (and some times mic pres) have all had a distinct effect on the tonality and stereo field. For example my John Hardy pres are transformed and seem to sound more honest than my cleaner Grace pre which seems to collapse the sense of depth a bit. This is actually the case with most of my transformered vs transformerless gear. I guess this could be distortion. Or it could be that transformers do some things better than transformerless designs, as well as add pleasant distortion.
Old 28th January 2015
  #73
Lives for gear
 

Hi
Valve preamps and 'low level' circuitry tends to be lower distortion, certainly less than say 0,05 percent unless they are being 'pushed'. It is only power amps and gear being 'hammered' that would get up to 5 percent.
The distortion is also scaled with the actual level, so again unless working hard the distortion will be pretty low.
Transformerless amplifiers can and usually are very low distortion although there are designs that are allowed to distort to give more 'flavour'.
When pushed hard, simple class 'A' designs tend to clip asymmetrically and often the distortion climbs with level. In contrast amplifiers with a high degree of negative feedback tend to be very clean and distortion free until all hell breaks loose when they 'clip'.
There are so many designs available that you cannot simply say 'valve good, transistor bad' as there are are examples of everything out there.
Transformers need to be used with the right 'conditions' to give their 'best' Impedances must be considered and operating levels must be suited to the design, both at high level (saturation) and at low level too.
Matt S
Old 29th January 2015
  #74
Deleted 6ccb844
Guest
Quote:
Originally Posted by IanBSC View Post
I'm actually surprised the distortion is that high. I am still not sure if the question of transparent vs colored isn't a false one. Some of the cleanest and most transparent converters I have used (and some times mic pres) have all had a distinct effect on the tonality and stereo field. For example my John Hardy pres are transformed and seem to sound more honest than my cleaner Grace pre which seems to collapse the sense of depth a bit. This is actually the case with most of my transformered vs transformerless gear. I guess this could be distortion. Or it could be that transformers do some things better than transformerless designs, as well as add pleasant distortion.
Yeah I suppose I should of explained a little more, I'm used to throwing numbers around in the lab. A solid triode amp would usually have around 0.001 - 3%THD at a certain level (watts).

Obviously the more you turn it up the more saturation you get, they say at around 4 - 6% total output you should avoid clipping and breakup with a tube amp. But amps are certainly not my field of expertise, it was interesting to mess around and do some testing with others who know more. So that's where the crank em thing comes from, probably everyone knew that but anyway.

So back to what I know about, audio interfaces..

Apparently all this THD+N that we don't want in our signal. We seem to now be chucking all that back in, (Also measurements by manufacturers can be a little skewed anyway, if I turn down the volume then the THD+N would be better) although the lower the THD+N apparently the more transparent and less noisy it is although it doesn't stop things from sounding different at those specs.

If I've got the choice I'd always have something like a Jensen adding THD and spurious frequencies back into the pot, than a piece of software trying to emulate it.. That's pretty much where this how good to pre-amps thing comes from, it's generally the opposite of what some are expecting..

It's becoming more and more common, slate VCC, saturation plugins, tape machines, mix busses, hardware compressors etc.

So the question is and believe me, this has been discussed in equipment circles more than the moon. Are we sure intermodulation /THD / Noise / Crosstalk and saturated compression isn't what we want?

The bigger question is, do we want our convertors adding spurious integer tonality back into it? Well only if it fits two criterion IMO, it has to be able to translate. No point in having beautiful sounding mixes from a DAC that sounds crap everywhere else and only on the ADC IF it's pleasing.

Finally, why not just use plugins or pre-amps to do this? Well if you think about a mixing console a large one with pre's / compressors / EQ's / faders.. There is potentially a lot of THD and the rest going on there, then some would add more by going to tape, then it would be mastered with more outboard.

So is a couple of plugins doing the same job (which every piece of equipment out there has the chance to sound slightly different) as the entire chain of an analogue suite?

Well probably not, but it gets better every day and the final question. Can conversion be too clean? IMO yes, it can and I hate to cop out at this point but it comes down to what works for you.

If the Burl is adding pieces to that clean chain you need, then it's perfect for the job.

Last edited by Deleted 6ccb844; 29th January 2015 at 12:31 AM..
Old 29th January 2015
  #75
Quote:
Originally Posted by Deleted 6ccb844 View Post
Are we sure intermodulation /THD / Noise / Crosstalk and saturated compression isn't what we want?
Yes, absolutely.

Allow me to explain why I feel so certain about it. Take these 5 simple aspects and try to create a cool sounding saturator/compressor out of them: It leads to garbage. None of these popular terms have any true beneficial musical effect.

That's why music production with a Sontec EQ sounds so much better than another EQ resulting in exactly the same specs. Cool sounding non-linearities have little to do with these 5 terms, they aren't particularly relevant in this case. On the other hand, it's perfectly possible to demonstrate that perfectly linear systems sound fantastic and represent the only option being able to achieve widest range of variety (saturated playback sounds.. ehmm.. saturated. Always. Reproduction of simple signals such as sines is impossible in such systems).

Back to topic. For monitoring systems there is absolutely no doubt that any change of the auditory information being transported is negative. It is restricting.

But opinions change with perspective, someone designing an ADA for a guitar effect pedal, or a mastering engineers looking for a certain effect in his loop will certainly have different opinions.

Last edited by FabienTDR; 29th January 2015 at 01:52 AM..
Old 29th January 2015
  #76
Deleted 6ccb844
Guest
Quote:
Originally Posted by FabienTDR View Post
Yes, absolutely.

Allow me to explain why I feel so certain about it. Take these 5 simple aspects and try to create a cool sounding saturator/compressor out of them: It leads to garbage. None of these popular terms have any true beneficial musical effect.

That's why music production with a Sontec EQ sounds so much better than another EQ resulting in exactly the same specs. Cool sounding non-linearities have little to do with these 5 terms, they aren't particularly relevant in this case. On the other hand, it's perfectly possible to demonstrate that perfectly linear systems sound fantastic and the only system being able to achieve widest range of variety (saturated playback sounds.. ehmm.. saturated. Always. Reproduction of simple signals such as sines is impossible in such systems).

Back to topic. For monitoring systems there is absolutely no doubt that any change of the auditory information being transported is negative. It is restricting.

But opinions change with perspective, someone designing an ADA for a guitar effect pedal, or a mastering engineers looking for a certain effect in his loop will certainly have different opinions.
THD and intermodulation is the epitome of "non-linear" that's the whole point of why we try to cut down on them so much for transparency and non spurious integers. No device is completely free of it.!

I agree noise is just garbage in some ways, but not completely. In audio telephony engineering we use such things as VAD to add comfort noise. So it doesn't confuse your brain and you start shouting hello.

Also why do we buy Tube amps, Pre-amps, plugins, consoles, tape machines? You name it engineers are adding it back!.

As I already said, nothing in the signal path should affect your listening. It has to translate, any product which causes variations that equate to miss-information can't be used in a professional line of work. But products like the Burl mothership adds a transformer behind the ADC, that will add all the gubins were talking about and people love it.

If there's only one thing I can take away from my entire career, there is no yes and no. It really just depends..
Old 29th January 2015
  #77
Lives for gear
 

sound itself is 'non-linear', so no need to chase the wrong specs in gear design.
Old 29th January 2015
  #78
Deleted 6ccb844
Guest
Quote:
Originally Posted by Timesaver800W View Post
sound itself is 'non-linear', so no need to chase the wrong specs in gear design.
That's great and all, but never would you want the outputs to your monitors adding frequencies which gives misinformation. So the point is for listening purposes to keep what was recorded the same as what you can hear.!

Were talking about adding some flavour to the ADC.

I should do a Shadow's guide to audio interfaces.
Old 29th January 2015
  #79
Lives for gear
 

i'm in total agreement. you don't want a converter to add frequencies, you don't want it to introduce time-smear. this is why converters with 'add-ons' sound better, it ofsets the time-smear.
Old 29th January 2015
  #80
Lives for gear
 

I used to have nice dCS and Lavry Gold and they were excellent (I am not current on the new products from these guys). But discrete designs are very expensive, and very time-consuming to troubleshoot. Maybe the slowdown is due to chip manufacturers slowing their development cycles? Also, the burr-brown and other solutions have come a long way, the technology has matured, and they have closed some of the gap with the esoterica.
Old 29th January 2015
  #81
Lives for gear
 
toledo3's Avatar
 

Quality of conversion has increased a great deal over the years in lower end products, to the point where they are often using the same chips - along with well made power supplies and other relevant parts - as higher end options. It's worth doing some investigation and real a/b-ing before blowing a ton of money on "high end" conversion.
Old 29th January 2015
  #82
Lives for gear
 

toledo3, I agree, that is exactly one of my points. Usually I use the converters in a Nagra VI. While not as great as in dCS or Lavry Gold it is still damn good. Of course, Nagra is not low end but still . . .
Old 29th January 2015
  #83
Deleted 6ccb844
Guest
Quote:
Originally Posted by toledo3 View Post
Quality of conversion has increased a great deal over the years in lower end products, to the point where they are often using the same chips - along with well made power supplies and other relevant parts - as higher end options. It's worth doing some investigation and real a/b-ing before blowing a ton of money on "high end" conversion.
Agree fully, I'm not going to name systems or we'll get into a flame war about it but some interfaces in the $700.00 range that's been released recently, from what I've seen of them are just as good as $3000.00 interfaces at the top three or four years ago.

IMO bang for buck right now (if you don't need much I/O) has to go to the Apogee Symphony. At the moment over here, they're going for the same price as a FireFace UFX. Sure if you need the I/O they quickly become expensive, but if not then there's not much that beats them in terms of specs and quality no matter how much money you have to spend.

But even now "low end interfaces" if it can be called that any more, are starting to nip they Symphony's heels. In a couple of years with the rate cheap interfaces are catching up I think high end is going to be a debateable term.

On a side note though, I have seen some "pro audio" pieces hanging about masquerading as quality items that is worse off than a Realtek HD.

Last edited by Deleted 6ccb844; 29th January 2015 at 08:59 PM..
Old 29th January 2015
  #84
Lives for gear
There is always distortion. We ought to look past it quantitively and address it qualitatively. If you have a very low % distortion, once a track gets loud that is going to be a significant portion of the sound - so it had better be a nice sounding distortion.
Old 30th January 2015
  #85
Lives for gear
 

I'm curious what the military, NASA, CERN, and other institutions that work with/measure sound in some capacity use in regards to conversion? My guess is converter technology has continued to develop as rapidly as the rest of digital technology. I'm just "thinking outside the box" since sound is not limited to music.
Old 30th January 2015
  #86
Lives for gear
 
Skip Burrows's Avatar
 

Quote:
Originally Posted by tekis View Post
I agree with this post. I've got a couple of friends in Ohio that make great rock records with 888's! There, I said it. When I listen to 'em, I don't hear anything that says, "If only they had better convertors..." In my own experience, I've always been fond of the way RADAR has sounded. They had me with RADAR II.
Yeah if I'm not mistaken Norah jones debut album come on over was recorded on a radar 1. That's a really nice sounding record. There are of course other examples but that one comes to mind.
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump